Fractal visuals have two build routes: domain repetition with raymarching for 3D lattices, or feedback zoom for a cheap 2D self-similar tunnel

The fractal style — self-similar detail across scales — can be built two ways. Route 1 (domain repetition): fold space with tiling-repeat (mod-space) around an SDF shape, then combine volumes via boolean-sdf; marching with raymarch-sdf produces infinite 3D lattices with genuine depth-layering. This is computationally heavier and GLSL-native. Route 2 (feedback zoom): run feedback-trail with a small scale-pulse per frame. Each frame is a slightly-zoomed copy of the previous; compounding produces an endless self-similar tunnel. This is cheaper, 2D, and live-friendly — Hydra’s native route. Adding radial-symmetry turns either approach into mandala-like patterns. Choosing between routes depends on context: 3D lattice needs GLSL; live improvisation favors feedback zoom.

Examples

Route 2 Hydra: src(o0).scale(1.01).rotate(0.01).out(o0) — every frame zooms in 1%, creating a compounding tunnel. Route 1 GLSL: space-fold with mod() around an sdBox, then raymarch.

Assessment

Name the two build routes for a fractal visual and state the primary trade-off between them. Which is preferred for live improvisation and why? Which requires GLSL?